Work around stability errors due to optional trailing commas #2126
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Optional trailing commas put by Black become magic trailing commas on another pass of the tool. Since they are influencing formatting around optional parentheses, on rare occasions the tool changes its mind in terms of putting parentheses or not. Ideally this would never be the case but sadly the decision to put optional parentheses or not (which looks at pre-existing "magic" trailing commas) is happening around the same time as the decision to put an optional trailing comma. Untangling the two proved to be impractically difficult. This shameful workaround uses the fact that the formatting instability introduced by magic trailing commas is deterministic: if the optional trailing comma becoming a pre-existing "magic" trailing comma changes formatting, the second pass becomes stable since there is no variable factor anymore on pass 3, 4, and so on. For most files, this will introduce no performance penalty since `--safe` is already re-formatting everything twice to ensure formatting stability. We're using this result and if all's good, the behavior is equivalent. If there is a difference, we treat the second result as the binding one, and check its sanity again.
JelleZijlstra
requested changes
Apr 25, 2021
| dst_contents = dst_contents_pass2 | ||
| assert_equivalent(src_contents, dst_contents, pass_num=2) | ||
| assert_stable(src_contents, dst_contents, mode=mode) | ||
| # Note: no need to explicitly call `assert_stable` if `dst_contents` was |
There was a problem hiding this comment.
We still need to call assert_equivalent though.
There was a problem hiding this comment.
We did, above in line 1018.
There was a problem hiding this comment.
So we do. Sorry for the noise!
JelleZijlstra
approved these changes
Apr 25, 2021
|
No. Those are still failing. I added new tests that pass after two calls to |
This was referenced Apr 25, 2021
Closed
Closed
This was referenced Apr 25, 2021
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Optional trailing commas put by Black become magic trailing commas on another pass of the tool. Since they are influencing formatting around optional parentheses, on rare occasions the tool changes its mind in terms of putting parentheses or not.
Ideally this would never be the case but sadly the decision to put optional parentheses or not (which looks at pre-existing "magic" trailing commas) is happening around the same time as the decision to put an optional trailing comma. Untangling the two proved to be impractically difficult.
This shameful workaround uses the fact that the formatting instability introduced by magic trailing commas is deterministic: if the optional trailing comma becoming a pre-existing "magic" trailing comma changes formatting, the second pass becomes stable since there is no variable factor anymore on pass 3, 4, and so on.
For most files, this will introduce no performance penalty since
--safeis already re-formatting everything twice to ensure formatting stability. We're using this result and if all's good, the behavior is equivalent. If there is a difference, we treat the second result as the binding one, and check its sanity again.